The scaling of internet delivery for radio broadcasters

A place to discuss Webcasting
Post Reply
User avatar
psmerdon
Posts: 9
Joined: Thu Feb 24, 2011 9:35 pm
Location: Melbourne, Australia

The scaling of internet delivery for radio broadcasters

Post by psmerdon » Sat Apr 30, 2011 12:05 am

I'm seeking thoughts on how internet radio will evolve for existing "over the air" broadcasters.

The trend to internet delivery of radio broadcasts is becoming noticeable ... with internet radio appliances, smartphones with apps like TuneIn Radio, and (coming soon) streaming to cars.
How do broadcasters adapt, without (to paraphrase Rupert Murdoch's print-media phrase) "swapping broadcast dollars for streaming cents"?

Certainly broadcasters have to cater to the streaming audience, to ensure that existing listeners don't drift off to all those thousands of other stations when they start using a streaming receiver.
Being local-centric with chat, traffic, and news springs to mind.

My worry is that streaming doesn't scale well economically.

Let's jump forward to the day when the Internet is at the point of becoming the primary audio information & entertainment delivery method (5 years, 10 years maybe?? or is that next month??).
Say a station had a peak listenership of 30,000 ppl (in morning drive)... I'm not an expert but I'd guess that's a mid-ranking metro station listenership here in Australia.
Does that number look generally realistic?

For the purpose of the exercise, let's assume that Internet Radio fragmentation has eroded a third of that audience.
Let's also assume that another third is still listening to your "over the air" signal, so you have to set up a stream to cater for 10,000 concurrent listeners (half your audience).
You'd want to build in some headroom to allow for audience growth, so let's make that 11,000.
What will it cost in bandwidth charges to deliver that number of individual TCP/IP or UDP streams?
You may be able to get it down to, what? 50 cents/month per stream?
i.e. $5,500/month, $66,000pa. That's big bux!

If we're optimistic, and say we'll retain all our 30,000 audience with a 50/50 split off-air/stream then that's 15,000 concurrent streams (plus headroom).
The more successful we are, the more expensive it gets.
(and maybe we can recover that extra cost - or maybe not...)

Today we don't have to buy more "transmitter" when the ratings go up.

But if we don't size our streams to be able to deliver to all comers, they'll drift off to another internet stream, not back to our off-air signal.
So they'll be lost to us, and we'll never be able to charge sufficiently for the advertising on the stream.

Remember, by this time we'll be looking to recover program creation costs plus a profit margin as well as the streaming costs - as the ability to subsidize the stream by using revenue from the main "over-the-air" channel is reduced as the audience there has been reduced (and only partially moved to the stream).

Is there a more efficient way to deliver streams in bulk?
My experience is that most stream listeners are local, and the rest are widely distributed - so content delivery network services (like Akami) probably won't help.

..or will streaming be likely to stay a niche service for a small "geekish" proportion of the audience?

I'd welcome your thoughts on this one, as I can't figure out how it's going to play out (sorry for the pun).
Sorry for being so long-winded.

Cheers,
Peter Smerdon
Melbourne, Australia

User avatar
BroadcastDoc
Site Admin
Posts: 2725
Joined: Tue Nov 13, 2007 3:34 pm
Location: Milwaukee, WI
Contact:

Re: The scaling of internet delivery for radio broadcasters

Post by BroadcastDoc » Sat Apr 30, 2011 12:38 pm

The good news is that it seems that for now, the streaming is simply being used to augment the terrestrial signal, so bandwidth costs will be kept fairly in check. I believe that advances in edge delivery and IP broadcast will reduce bandwidth and usage and push the delivery out to the "last mile" and that the cost concern isn't the bandwidth - it's the royalties. If you had 30,000 people listening, and you're paying $.0025 per person per song (2015 Royalty board rates) and playing ten songs an hour, that's $750/hr that you have to pay in royalties. Ouch!
Christopher "Doc" Tarr CSRE, DRB, AMD, CBNE
Help support the Virtual Engineer, use our 1&1 Affiliate link if you need good, cheap hosting.
Virtual Engineer. The Broadcast Engineering discussion forum

User avatar
Deep Thought
Posts: 3168
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: The scaling of internet delivery for radio broadcasters

Post by Deep Thought » Sat Apr 30, 2011 12:50 pm

As long as the RIAA controls SoundExchange and charges obscene performance royalty rates for music on the Internet, this is a non-starter. The royalties will outpace the streaming cost. Unlike OTA broadcasting there is a marginal cost for each additional listener, and those costs are difficult to recoup let alone make a profit over. When you start talking numbers like those of a high-rated major market station, it does not make any sense at all. If you translate a 33,000 person AQH audience to the web, every song you play would cost you $56.01 at current rates ($0.0017 per stream per track). At 10 songs per hour, that is $13,464 per day, $4,914,360 per year! And that does not count the other fees you are already paying to play music over the air. Even a medium market station with a 3,000 person AQH would find this prohibitive (almost $500K) and a small market station with 300, forget it. There is a strong financial disincentive to moving listeners to the Internet, and those who are already on it have audience numbers nothing like what even a mildly successful middle-market radio station does, and they still struggle to make ends meet.

Since we're unlikely to see anything approaching universal service in wireless data, listening to "radio" over the Internet in a mobile environment will remain difficult to do away from urban areas and major roadways due to coverage issues and it will become expensive to do where there *is* coverage since the wireless data players are all moving to a pay-per-byte model. Streaming is a data-intensive use and AT&T, Verizon et. al. are salivating at the prospect of $billions of new revenue from it. Plus, the technology is still unreliable, resulting in a lot of drop-outs, reconnects and no-connects as you move from one cell to another.

Fixed-location Internet "radio" fares a lot better but again the bandwidth providers are all moving to a pay-per-byte model (mostly in reaction to video, which competes with their established pay-TV businesses) which also affects streaming audio.

Multicast can help alleviate some of the transport costs, and there are schemes like Octoshape which can set up a peer-to-peer distribution network, but you're still on the hook for those performance royalties and it does not help with the "last mile" issues I mentioned above.

At this point streaming is more of an adjunct for most stations, reaching those who are trying to listen in poor reception environments or are away from home. I don't know any stations making money with it, and most are piss-poor quality due to the expense involved with high-rate streams and/or due to being locked into proprietary "players" which display banner ads.

Radio stations need to look at their respective markets and decide what it is worth. The claims of "stream or die" are from those who stand to make money off the streaming, or by creating controversy which in turn drives blog hits, so you can safely ignore them. It is important to have a competent presence on the web, something many stations fail miserably at. Those streaming dollars would be much better spent developing that.

(I see Doc snuck one in while I was typing this, and he raises the same point...)
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

am1430
Posts: 24
Joined: Fri May 21, 2010 7:36 am

Re: The scaling of internet delivery for radio broadcasters

Post by am1430 » Mon May 02, 2011 4:41 am

interesting thread on streaming since my wife and i are considering streaming especially for our nighttime programming with high school sports, etc. i guess my question would be if the royalty rates are so high, how can a company like Pandora make it..even Chevrolet is announcing they will have some type of receiver that will get Pandora..I know Pandora is doing some ad sales..but is it really worth it for the clients since there are so many streaming stations?

User avatar
Deep Thought
Posts: 3168
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: The scaling of internet delivery for radio broadcasters

Post by Deep Thought » Mon May 02, 2011 3:38 pm

Tim Westergren, the owner of Pandora, is in cahoots with the RIAA and negotiated special rates for himself in exchange for selling out his fellow streamers and broadcast stations. Now he's the mouthpiece for the RIAA in this royalty battle. Seriously, if you care anything at all about the future of Internet streaming or even music on the radio you need to rethink supporting him and Pandora in any way.

/climbs off soapbox.
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

am1430
Posts: 24
Joined: Fri May 21, 2010 7:36 am

Re: The scaling of internet delivery for radio broadcasters

Post by am1430 » Tue May 03, 2011 4:14 am

oh trust me deep..im not a fan of pandora..i think they are a threat to local radio.

User avatar
psmerdon
Posts: 9
Joined: Thu Feb 24, 2011 9:35 pm
Location: Melbourne, Australia

Re: The scaling of internet delivery for radio broadcasters

Post by psmerdon » Mon May 09, 2011 4:48 pm

Point of reference - I'm not in the USA - I'm in Australia, so I don't know much about the "streaming rights" battles over there
OK, so for the USA the music rights cost issue is hampering the growth of streaming by existing broadcasters.
How does that differ from non-broadcast music streamers other than Pandora - which from previous posts looks like it has a sweetheart deal?
Here in Oz, the commercial radio industry body (Commercial Radio Australia) has a deal with the music licensing body (APRA/AMCOS) for streaming rights fees. I don't know what it is, but I haven't heard of any complaints so the industry must be reasonably satisfied.

Deep Thought says "There is a strong financial disincentive to moving listeners to the Internet..."
I'm not sure you're moving them there - many of them may be moving there regardless of what you do.
If your listeners are gradually moving online, shouldn't you be waiting there for them?

I was more looking for commentary on the issue of poor scalability of streaming vs over-the-air.
Deep Thought, you mention Multicast and services like Octoshape (which, to me looks like a Content Delivery Network like Akami).
My observations from the support calls and comments I have gotten for stations I've streamed is that the vast majority of streaming listeners are in the station's primary "catchment area". During business hours most are in central city offices - using streaming 'cos their radios don't work inside the high-rise offices. As the smartphone streaming apps and in-car streaming devices increase their penetration, this majority's listening will expand to the drive into and home from work - along major commuting corridors.
What's the CDN service's benefit here, compared to delivering the stream directly thru my local streaming servers?
How about Multicast? I'd appreciate a heads-up on the benefits, and what's needed to implement that.
I did find this interesting thread in the Aussie Whirlpool forum "ISPs in Australia" (an excellent set of forums!):
http://forums.whirlpool.net.au/archive/1172682
Is this accurate in your experience??

Thanks all for the interesting comments.
Please keep them coming. This topic is not going to go away.

Cheers,

Peter Smerdon
Peter Smerdon
Melbourne, Australia

User avatar
Deep Thought
Posts: 3168
Joined: Thu Mar 20, 2008 9:23 am
Location: La Grange, IL
Contact:

Re: The scaling of internet delivery for radio broadcasters

Post by Deep Thought » Mon May 09, 2011 6:55 pm

Peter, you may have misread what I wrote, which is not your fault. My intent saying that there is a financial disincentive to move listeners to Internet streams is precisely what I said: The cost is far higher than any benefit here in the USA. "Being ready" is nice but until some sanity is restored in the performance rights arena here encouraging stream listening in lieu of OTA is akin the old joke...we lose money on every sale but make it up in volume! Read what I posted above about the cost to the station.

Expansion of online listening here is also being throttled by wireless phone companies imposing punitive usage charges on anything but the puniest data flow. Would you pay $100/month to listen on your phone when you can listen for free on the radio? That's where we are headed. Rates are going up, fast.

Talk radio has a different set of rules since they usually only use snippets of musical works as bumpers, etc. and (unless you are Rush Limbaugh) usually obtains permission beforehand if they are used as themes. Commercial beds are similarly exempt from performance royalties since they are covered under the ASCAP/BMI licenses or are purchased as licensed packages and are not considered "performances".

Octoshape is mainly a peer-to-peer streaming network, not a client-server arrangement. You can read more about it here: http://www.octoshape.com/
Mark Mueller • Mueller Broadcast Design • La Grange, IL • http://www.muellerbroadcastdesign.com

User avatar
psmerdon
Posts: 9
Joined: Thu Feb 24, 2011 9:35 pm
Location: Melbourne, Australia

Re: The scaling of internet delivery for radio broadcasters

Post by psmerdon » Tue May 10, 2011 4:25 pm

Thanks for the response, Deep Thought.
I read your post about the costs - but couldn't believe those charges could be sustained.
Surely it's not in the interests of the music industry to impose such costs as to prevent mass streaming of their product.
I'd have thought somebody would be negotiating volume licensing, using eg. your cost example to demonstrate the flaw in their model.
Or have they dug their heels in - and want big (extortionate) fees from a lot of little player ... with, say, a few dozen to hundreds of listeners, than to allow the growth of big streamers.
Except Pandora - that situation seems a bit smelly if what you say is true.

Anyway I was hoping for a more technical discussion on how we sensibly scale the streaming model to better suit 10 or 20 thousand listeners.
I'll look further into Octoshape. Would it help the issue of 20thousand listeners all in the same city using 20k individual streams and bandwidth from my server?

Cheers,
PeterS
Peter Smerdon
Melbourne, Australia

Kent T
Posts: 101
Joined: Mon Dec 17, 2012 7:29 pm

Re: The scaling of internet delivery for radio broadcasters

Post by Kent T » Thu Sep 05, 2013 5:56 pm

Until SoundExchange and RIAA Streaming Royalty Greed is eliminated. Heavily capped cellular data plans are eliminated, and broadband is everywhere at reasonable speeds, I do not envision Internet Delivery as a main method. Most cellular data carriers have very onerous data use policies and limits, and there is not the bandwidth in most areas outside major cities for Internet delivery to be reliable and a broadcaster's bread and butter.

TPT
Posts: 640
Joined: Mon Dec 03, 2007 3:18 pm
Location: St. Marys, WV

Re: The scaling of internet delivery for radio broadcasters

Post by TPT » Sat Sep 07, 2013 6:35 am

We stream sports on two of our stations and full-time stream another station. There is a basic rate available for small stations (we don't have very many average listener hours or whatever they're called) so cost is about a grand a year for the full time streaming. Streaming the games is about $75/month which we cover with a local sponsor.

Assuming a large audience could be created there is another problem--advertisers. Don't know about Australia, but here many national advertisers run different campaigns in different parts of the country. McDonald's, for example. Our local McDonald's guy owns two stores--one in the small town where our main station is, one 25 miles away where I have another station. There are two different agencies buying time for those two different McDonalds--often with two different promotions. Which causes confusion when our one station (a B1--25kw) covers both areas. Now throw in streaming, with no geographic limits.....

As for local direct, very difficult sell to convince them on the merits of streaming (unless it's thrown in for free). Sell streaming alone? Fuggadiaboutit!

Kent T
Posts: 101
Joined: Mon Dec 17, 2012 7:29 pm

Re: The scaling of internet delivery for radio broadcasters

Post by Kent T » Sat Sep 07, 2013 10:11 am

At home, usually streaming is not bad. Unless you are in a very rural area. But cellular data is where the infrastructure is not there yet. And the Cellular data and satellite internet data limits are really chintzy. Cable, Fiber Optic, and DSL is not as onerous. At work, stateful firewalls are really beginning to be an issue.

Baylink
Posts: 438
Joined: Wed Dec 15, 2010 3:21 pm

Re: The scaling of internet delivery for radio broadcasters

Post by Baylink » Mon Sep 23, 2013 11:31 am

To clarify one point: Octoshape isn't Akamai, it's *BitTorrent*. It uses your listeners PCs to redistribute the signal to other listeners.

But in general no: regardless of how many of your listeners move to streaming ,it is not technically or financially scalable with the current workaround technologies we're using to do it. Wide-Area Multicast, properly implemented at all layers of the internet will be necessarily technically (hint: it ain't yet), and the layer 8 9 and 10 problems may be insoluble (people, money, lawyers).

Post Reply